datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
amd-nrtransformers_pr_citransformers_daily_ciAM-DeepSeek-Distilled-40MFor more open-source datasets, models, and methodologies, please visit our GitHub repository and paper: DeepDistill: Enhancing LLM Reasoning Capabilities via Large-Scale Difficulty-Graded Data Training.
Due to certain constraints, we are only able to open-source a subset of the complete dataset.
Model Training Performance based on our complete dataset
On AIME 2024, our 72B model achieved a score of 79.2 using only supervised fine-tuning (SFT). The 32B model reached 75.8 and… See the full description on the dataset page: https://huggingface.co/datasets/a-m-team/AM-DeepSeek-Distilled-40M.AM-DeepSeek-R1-Distilled-1.4MFor more open-source datasets, models, and methodologies, please visit our GitHub repository.
AM-DeepSeek-R1-Distilled-1.4M is a large-scale general reasoning task dataset composed of
high-quality and challenging reasoning problems. These problems are collected from numerous
open-source datasets, semantically deduplicated, and cleaned to eliminate test set contamination.
All responses in the dataset are distilled from the reasoning model (mostly DeepSeek-R1) and have undergone
rigorous… See the full description on the dataset page: https://huggingface.co/datasets/a-m-team/AM-DeepSeek-R1-Distilled-1.4M.Instella-Long
Instella-Long
The Instella-Long dataset is a collection of pre-training and instruction following data that is used to train Instella-3B-Long-Instruct. The pre-training data is sourced from Prolong. For the SFT data, we use public datasets: Ultrachat 200K, OpenMathinstruct-2, Tülu-3 Instruction Following, and MMLU auxiliary train set. In addition, we generate synthetic long instruction data using documents of the books and arxiv from our pre-training corpus and the dclm subset from… See the full description on the dataset page: https://huggingface.co/datasets/amd/Instella-Long.ReasonLite-Dataset
GitHub |
Dataset |
Blog
ReasonLite is an ultra-lightweight math reasoning model. With only 0.6B parameters, it leverages high-quality data distillation to achieve performance comparable to models over 10× its size, such as Qwen3-8B, reaching 75.2 on AIME24 and extending the scaling law of small models.
🔥 Best-performing 0.6B math reasoning model
🔓 Fully open-source — weights, scripts, datasets, synthesis pipeline⚙️ Distilled in two stages to balance efficiency and high… See the full description on the dataset page: https://huggingface.co/datasets/amd/ReasonLite-Dataset.Micro-World-MC-DatasetMicro-World is an interactive world model developed by AMD. We leverage the Minecraft API MineDojo to collect game data, as it enables us to obtain diverse biomes as well as varying weather and lighting conditions—properties that are desirable for improving generalization to real-world scenarios.
For the action space, we record keyboard controls including W (forward), A (left), S (backward), D (right), Ctrl (sprint), Shift (sneak), Space (jump), along with mouse movement. By randomly sampling… See the full description on the dataset page: https://huggingface.co/datasets/amd/Micro-World-MC-Dataset.AM-DeepSeek-R1-0528-Distilled
📘 Dataset Summary
This dataset is a high-quality reasoning corpus distilled from DeepSeek-R1-0528, an improved version of the DeepSeek-R1 large language model. Compared to its initial release, DeepSeek-R1-0528 demonstrates significant advances in reasoning, instruction following, and multi-turn dialogue. Motivated by these improvements, we collected and distilled a diverse set of 2.6 million queries across multiple domains, using DeepSeek-R1-0528 as the teacher.
A notable… See the full description on the dataset page: https://huggingface.co/datasets/a-m-team/AM-DeepSeek-R1-0528-Distilled.AM-DeepSeek-R1-Distilled-1.4MAIG-Datasets
AMD AIG GPU Kernel Datasets
AMD AIG-Datasets is a collection of GPU-kernel generation, translation, optimization,
and ROCm-library supervision data. It contains PyTorch/CUDA-to-HIP, HIP-to-HIP,
PyTorch-to-Triton, and production-grounded rocBLAS/rocSOLVER entries, together with
metadata, samples, conversion utilities, and reproducible evaluation tools.
The repository is organized into versioned releases. New training and evaluation
workflows should use the unified-schema datasets… See the full description on the dataset page: https://huggingface.co/datasets/amd/AIG-Datasets.Cot-Drop
LuminaSFT
LuminaSFT is a synthetic SFT dataset suite specifically designed to improve both general-purpose and task-specific SLMs. LuminaSFT consists of multiple curated splits that target diverse capabilities:
UltraChat200K-DeepSeek - A regenerated base SFT dataset for broad instruction following.
InstructGPT-NaturalQA and InstructGPT-TriviaQA - Factual question answering datasets to strengthen knowledge recall and answer accuracy.
CoT-Drop - A reading comprehension dataset with… See the full description on the dataset page: https://huggingface.co/datasets/amd/Cot-Drop.AM-DeepSeek-R1-Distilled-1.4M-EnglishAudioJailbreak
Audio Jailbreak: An Open Comprehensive Benchmark for Jailbreaking Large Audio-Language Models
AudioJailbreak is a benchmark framework specifically designed for evaluating the security of Audio Language Models (Audio LLMs). This project tests model defenses against malicious requests through various audio perturbation techniques.Note: This project aims to improve the security of audio language models. Researchers should use this tool responsibly.
📋 Table of… See the full description on the dataset page: https://huggingface.co/datasets/amd1234567/AudioJailbreak.AM-DeepSeek-R1-0528-Distilled-with-Systembert-base-uncased-2022_tokenized_datasetxbmu_amdo31
Dataset Card for [XBMU-AMDO31]
Dataset Summary
XBMU-AMDO31 dataset is a speech recognition corpus of Amdo Tibetan dialect. The open source corpus contains 31 hours of speech data and resources related to build speech recognition systems, including transcribed texts and a Tibetan pronunciation dictionary.
Supported Tasks and Leaderboards
automatic-speech-recognition: The dataset can be used to train a model for Amdo Tibetan Automatic Speech Recognition (ASR). It… See the full description on the dataset page: https://huggingface.co/datasets/syzym/xbmu_amdo31.SAND-Post-Training-Dataset
SAND-Post-Training-Dataset: High-Quality Synthetic Reasoning Dataset Built with AMD GPUs
Dataset Summary
We introduce the SAND-Post-Training-Dataset, a high-quality synthetic reasoning dataset for mathematics and science built entirely using a synthetic data pipeline running on the AMD ROCm™ stack and AMD Instinct™ MI325 GPUs.
This dataset prioritizes difficulty and novelty over volume, demonstrating that high-difficulty synthetic data can elevate… See the full description on the dataset page: https://huggingface.co/datasets/amd/SAND-Post-Training-Dataset.SAND-MATH
SAND-Math: A Synthetic Dataset of Difficult Problems to Elevate LLM Math Performance
📃 Paper | 🤗 Dataset
SAND-Math (Synthetic Augmented Novel and Difficult Mathematics) is a high-quality, high-difficulty dataset of mathematics problems and solutions. It is generated using a novel pipeline that addresses the critical bottleneck of scarce, high-difficulty training data for mathematical Large Language Models (LLMs).
Key Features
Novel Problem Generation: Problems are… See the full description on the dataset page: https://huggingface.co/datasets/amd/SAND-MATH.flores_plushackathon_amd_mission2_black_sort_fixedThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 145,
"total_frames": 44304,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:145"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/giacomoran/hackathon_amd_mission2_black_sort_fixed.amdpilot-lora-sft-dataset
AMDPilot LoRA SFT Dataset
SFT training data for fine-tuning LLMs on AMD GPU debugging, optimization, and kernel engineering tasks. Each example is a multi-turn conversation in OpenAI messages format with tool-use annotations.
Usage
from datasets import load_dataset
# Load a specific version
ds = load_dataset("JinnP/amdpilot-lora-sft-dataset", "v5_2")
# Load a specific view
ds = load_dataset("JinnP/amdpilot-lora-sft-dataset", "v5_2_chunks")
# Available configs: v4, v5… See the full description on the dataset page: https://huggingface.co/datasets/JinnP/amdpilot-lora-sft-dataset.InstructGpt-educational
LuminaSFT
LuminaSFT is a synthetic SFT dataset suite specifically designed to improve both general-purpose and task-specific SLMs. LuminaSFT consists of multiple curated splits that target diverse capabilities:
UltraChat200K-DeepSeek - A regenerated base SFT dataset for broad instruction following.
InstructGPT-NaturalQA and InstructGPT-TriviaQA - Factual question answering datasets to strengthen knowledge recall and answer accuracy.
CoT-Drop - A reading comprehension dataset with… See the full description on the dataset page: https://huggingface.co/datasets/amd/InstructGpt-educational.hackathon_amd_mission2_black_flipThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 150,
"total_frames": 49863,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:150"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/giacomoran/hackathon_amd_mission2_black_flip.AM-DeepSeek-R1-Distilled-1.4M-am_0.5MAM-DeepSeek-R1-Filtered-Math-Code
AM DeepSeek R1 Filtered Math and Code
This repository publishes reproducible training subsets derived from
a-m-team/AM-DeepSeek-R1-Distilled-1.4M
at pinned revision 53531c06634904118a2dcd83961918c4d69d1cdf.
The source dataset and this derived release use CC BY-NC 4.0. Commercial use is
not permitted by that license. Preserve attribution and review the upstream
dataset card before use.
Contents
Config / split
Records
Bytes
SHA-256
math / train
111,657
2… See the full description on the dataset page: https://huggingface.co/datasets/chhao/AM-DeepSeek-R1-Filtered-Math-Code.hackathon_amd_mission2_blue_pickThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 151,
"total_frames": 36913,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:151"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/giacomoran/hackathon_amd_mission2_blue_pick.UltraChat200K-regenerated
LuminaSFT
LuminaSFT is a synthetic SFT dataset suite specifically designed to improve both general-purpose and task-specific SLMs. LuminaSFT consists of multiple curated splits that target diverse capabilities:
UltraChat200K-DeepSeek - A regenerated base SFT dataset for broad instruction following.
InstructGPT-NaturalQA and InstructGPT-TriviaQA - Factual question answering datasets to strengthen knowledge recall and answer accuracy.
CoT-Drop - A reading comprehension dataset with… See the full description on the dataset page: https://huggingface.co/datasets/amd/UltraChat200K-regenerated.Instella-GSM8K-synthetic
Instella-GSM8K-synthetic
The Instella-GSM8K-synthetic dataset was used in the second stage pre-training of Instella-3B model, which was trained on top of the Instella-3B-Stage1 model.
This synthetic dataset was generated using the training set of GSM8k dataset, where we first used Qwen2.5-72B-Instruct to
Abstract numerical values as function parameters and generate a Python program to solve the math question.
Identify and replace numerical values in the existing question with… See the full description on the dataset page: https://huggingface.co/datasets/amd/Instella-GSM8K-synthetic.telephony-amd-dataset
Telephony AMD (Answering Machine Detection) Dataset
Overview
A multilingual 4-class telephony audio classification dataset for training streaming Answering Machine Detection models. Contains real human speech (PolyAI/MINDS14) mixed with TTS-generated audio (Microsoft Neural TTS / edge-tts) across English, French, Spanish, and German.
Key design principle: Voicemail greetings are recorded by real humans and sound acoustically identical to live speech. This dataset includes… See the full description on the dataset page: https://huggingface.co/datasets/AbijahKaj/telephony-amd-dataset.